We are trying something a little different for our first Columbus Futurists meeting of the new year. There are 3 sections below:
- The Idea;
- To Do
Initial participants and their forecasts are shown in the spreadsheet (pdf) linked here and below.
Some of us have recently read Philip Tetlock’s new book, Superforecasting: The Art and Science of Prediction. A well-known forecaster in his own right, Tetlock describes an experiment – contest, if you will – involving pitting professional forecasters against a group of volunteer amateur forecasters. The pros had had long careers and track records in the business. The amateurs had little to no experience, and only accidental training or familiarity with research or analytical techniques, if any.
The object was to make forecasts about specific changes or events that would happen in set time frames. The forecasters could research as much or as little as they wished, and could update their forecasts as often or not – as they wished. At specific times – the accuracy of the forecasts – in terms of how close the forecast was to the actual objective state of the topic – was taken. The accuracy was calculated by a standard algorithm which created an index of accuracy for each forecaster. See below for a little more background – but the amateurs did well.
So – with that introduction – we would like to try something similar, less elaborate but in the same realm. Columbus Futurists are all very bright, well-informed, curious people who pay attention to a broad range of issues around them. So we thought it might be fun to challenge everyone to pick a topic or two or three, make some forecasts about their futures in a somewhat structured way – and then see how things turn out – perhaps next January, 2017?
We were thinking that it would be wise – both to focus our limited membership’s energies, and to facilitate tracking the results over time – to agree on a handful of questions that everyone would tackle. And then we will track the results.
From a practical perspective, and for ease of evaluation, we might want to start with relatively small, time compressed events or trends – so we can keep interest up at the beginning and make adjustments over time if we wish. Perhaps if the idea takes hold – we could find a way to start our own small “decision market” later in the year to follow longer term issues? Or maybe it becomes something of a new year’s tradition.
Here are the questions with which we propose to start:
- Politics: Who will win their respective parties nominations for the presidential election?
- Republican Convention – ending July 21, 2016, Cleveland OH
- Democratic Convention – ending July 28, Philadelphia, PA
- Politics: Who will win the Ohio primaries (Dem & Rep) – March 15, 2016?
- Economy: What will the unemployment rate be on December 1, 2015?
- For the state of Ohio
- For the US as a whole?
- What will the Dow Jones Industrial Average be on December 1, 2015?
- Geopolitics: What is the likelihood of a major US troop deployment (5000 or more)
- Global climate: What is the likelihood of a 50% or higher remission in drought in the American West, focus on California?
Tetlock’s new book, Superforecasting: The Art and Science of Prediction wouldn’t be much of a book if the pros smoked the amateurs. Turns out the volunteers did quite well for themselves. A great deal was made of finding an objective measurement to compare results intra- and inter-team. And, while there were stars and also-rans on both sides, there were amateurs who did spectacularly well.
I won’t go into all the detail – the story as it evolves is one of the pleasures of the book. But it is a good read, and contains a lot of valuable insight – especially for us amateurs, about how different people with different perspectives go abut dealing with complicated challenges.
In order to keep track of success, the issues around which the research and forecasting was done were kept to rather discrete, fairly objective topics. They covered a broad range of disciplines, and with consultation, each participant was assigned the issues they were to follow. In the end, however, the comparisons were made on the basis of relative success: who got closest to the reality? – rather than which issues came to what end.
- Decide on the topics you wish to follow
- Send them to myself or the both David and I – by Tues Jan 19 – or bring them to the meeting
For each forecast:
- Prediction must include objective change in conditions, a measurable result
- Prediction must be specific, a single data point (not a list or a range)
- Prediction can include a percentage of likelihood from the predictor
Our collation of the submissions will show all of this information accompanying each prediction.
The next meeting of the Columbus Futurists is Jan 21. We apologize for the brief time for preparation (let’s blame the holidays!).
But we want to be as inclusive as possible – so we wish to encourage you – even if you do not plan to attend the meeting – or even if you do – to send us your forecast by email by Tuesday, January 19. This way you do not have to be present to participate. We will try to collate all the forecasts by the time of the meeting on Thursday, and then we can discuss how we will proceed.
The forecasts will be published on the Columbus Futurists web site. You can modify your forecast whenever you like – but we will encourage at least one major checkpoint mid-year – in July.
There are no doubt details to be worked out and we can discuss them on the 21st. Meantime, take some time to look into the issues you wish to pursue. Come up with a forecast that seems reasonable to you (remember we are looking for the closest result as of the target date mentioned in each topic statement).
Both David and I are excited about the possibilities, and hope you see the potential for some fun and some learning. Feel free to reply to both of us with any feedback.
In addition to Tetlock’s book, here’s an interesting new tool – called Guesstimate – that purports to help with calculations of uncertainty. Read the PR and check out the link to Guesstimate if you wish to try it.
And see you soon!
Best – Rich Bowers
David Staley, President